On more than one occasion during my time at GitHub, teams were formed to tackle onboarding, engagement, and retention concerns. With new user experience being close to my heart, I put my hand up each time the opportunity came up.

Due to the exploratory nature of working on user growth and onboarding efforts, my time on these teams involved partnering closely with research, data, and analytics teams. This is an area that GitHub hadn’t historically invested in to a large degree, so much of the work was laying the foundations—building up our internal tooling for surveys and processes for collecting feedback, comprehensive instrumentation of key workflows, and running experiments to determine baselines for behaviour we could rely on later when evaluating opportunities for optimisation.

Some of the earlier successes we saw on this team were a redesign of the "blankslate" welcome page that boosted 90-day retention by ~30%, the addition of email verification functionality to the registration process as a way to improve activation, and the development and implementation of a custom-built survey framework which would enable us to run all kinds of experiments going forward in reliable and flexible ways—collecting data to power a number of interesting findings (e.g. did you know that 70% of people signing up to GitHub each week don’t know Git? Fascinating huh!) and empowering teams across the company to quickly explore research as well.

Even among teams where qualitative research capacities exist, it’s an unfortunate reality that way too many teams try to iterate on product without proper instrumentation, analytics, and data-driven insights. Sometimes you only learn where things need to go by suspending your preconceived biases and opinions, and take the time to look at the hard truths revealed in data by concrete behaviour.